agentleFS
tanod-labstanod-labs/tanod-mcp/skills/pdf-tables/SKILL.md

pdf-tables

Skill1 starsChanged 34 days agoMITView on GitHub

Extracts the tables of a text PDF as JSON, CSV or XLSX with Tanod Documents. Use when an agent must pull tables out of a PDF for analysis.

The file

Files
SKILL.md

PDF tables (Tanod Documents)

POST /v1/pdf/tables returns every table of the selected pages (ruled tables and tables found from text alignment). Price: USD 0.01 per call. Free allowance: none (PDF tables are never free).

Use when

  • a PDF holds tables that must become JSON, CSV or a spreadsheet
  • rows and columns must be kept, not flattened to text

Use instead

  • whole document as Markdown: document-to-markdown
  • scanned PDF (422 no_text_layer): run the ocr skill first
  • a bank statement with reconciliation to OFX/QIF: bank-statement

Request

JSON body on POST https://tanod.dev/v1/pdf/tables. Files are sent as a URL or base64 in JSON; these routes do not take multipart/form-data.

  • url: public http(s) URL of the PDF, at most 20 MB; or file_base64: the PDF as base64, at most 10 MB decoded.
  • pages (optional): N, N-M, N-, -M, last, comma separated, e.g. 1-5,9; at most 200 pages per call; default all.
  • format: json (default), csv (one file per table) or xlsx (one workbook, a sheet per table).
  • output: base64 (default, file inline) or url (24-hour download link).

Example body: {"url": "https://tanod.dev/samples/tables-sample.pdf"}

POST /v1/pdf/tables HTTP/1.1
Host: tanod.dev
content-type: application/json

{"url": "https://tanod.dev/samples/tables-sample.pdf"}

MCP alternative: connect to https://tanod.dev/mcp and call the tool extract_pdf_tables. No free allowance on this tool.

Payment

An unpaid request returns HTTP 402 with an x402 v2 challenge (USDC on Base, Polygon or Solana) and an MPP/Tempo WWW-Authenticate: Payment challenge. Pay it and repeat the request with the payment attached. No account or API key. Input errors (413, 415, 422) and worker failures (429, 502, 503, 504) are not charged.

Response

tables (each with name, page, bbox, method lines|text, header, rows of strings), table_count, pages_without_text, pages_processed, warnings; for csv/xlsx the file comes back in file or files.

Limits

  • PDF up to 500 pages, 10 MB as base64 (20 MB via url), at most 200 selected pages per call (422 too_many_pages).
  • Text PDFs only; pages with no table give 422 no_tables_found.
  • More than 5 MB of JSON (15 MB of CSV/XLSX) is a 413 output_too_large.
  • Encrypted input is a 422 pdf_encrypted.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.